5 – Measurement and Scaling Techniques

Introduction

Measurement assigns numbers or other symbols to properties of things measured by pre-set principles. A concept (or construct) is a broad idea about a group of objects, qualities, events, or processes. Age, gender, number of children, education, and wealth are solid constructs. Brand loyalty, personality, channel power, and satisfaction are relatively abstract constructs. Scaling is creating a continuum on which measured things are positioned. A scale is a quantitative measure—a grouping of elements organized progressively according to value or magnitude. The goal is to quantify an item, person, or event’s position on the scaling continuum.

5.1 Measurement Scales: Tools of Sound Measurement

These are divided into four types of scales, namely:

• Nominal scale
• Ordinal scale
• Intervals scale
• Ratios scale

 5.1.1 Nominal Scale

Numbers are used to identify the things on this scale. For instance, university registration numbers, such as the numbers on their jerseys, are issued to students.

In this style of scaling, the objective of marking numbers, symbols, labels, and so on is not to establish order but rather to place labels to identify occurrences and count the objects and subjects. This measurement scale categorises individuals, businesses, products, brands, or other things into categories with no inherent order. It is frequently referred to as a categorical scale. It is a classification system that does not place the entity on a continuum. It entails a simple count of the frequency of cases assigned to the various categories, and if required, nominal numbers can be assigned to label each category.

Characteristics

• It does not have an arithmetic origin.
• It demonstrates no orderly or distant relationship.
• It distinguishes between items by categorising them.

Use: This scale is commonly used in surveys and ex-post-facto research.

For instance, have you ever been to Bangalore?

Yes -1, No -2

‘Yes’ is coded as ‘One’ and ‘No’ as ‘Two.’ The numeric code associated with the responses is irrelevant and serves just as an identification. Respondents’ answers will not change if the numbers are adjusted to one for ‘No’ and two for ‘Yes.’ The numbers used in nominal scales are merely for counting.

Telephone numbers are an example of a nominal scale, with one number allocated to each subscriber. The purpose of utilising a nominal scale is to ensure that no two people or objects are assigned the same number. Bus route numbers are another example of nominal scale.

“What age are you?” This is an illustration of a nominal scale. “Can you tell me your PAN Card number?”

We arrange the volumes in the library on a nominal scale, subject by subject and author by author.

• There is no ranking system.
• There is no mathematical operation that can be performed.
• Statistical implications—The standard deviation and mean cannot be computed. The model can be expressed in various ways.

5.1.2 Ordinal Scale (Ranking Scale)

In most market research studies, the ordinal scale is employed for ranking. Ordinal scales determine consumer views, preferences, and so on. For example, respondents could be given a list of brands that might be suitable and asked to rank them on an ordinal scale:

• Lux
• Liril
• Cinthol
• Lifebuoy
• Hamam

Rank	Item	Number  of respondents
I	Cinthol	150
II	Liril	300
III	Hamam	250
IV	Lux	200
V	Lifebuoy	100
Total		1,000

In the preceding example, II is the mode, and III is the median.

Statistical implications: The mode and median can be calculated.

In market research, we frequently ask respondents to rank goods, such as “A soft drink, depending on flavour or colour.” In such a circumstance, the ordinal scale is applied.

Rank the following microwave oven attributes on a 1–5 scale in order of importance:

Attributes	Rank
A)    Company Image	5
B)  Functions	3
C)    Price	2
D)    Comfort	1
E)  Design	4

The ordinal scale organizes things. In qualitative research, rank ordering ranks characteristic units from most important to least important.

Characteristics

• The ordinal scale ranks items from highest to lowest.
• These scales are not given in absolute terms.
• The difference in rank between adjacent ranks is not always equal.
• The median is used to calculate central tendency.
• The percentile or quartile is used to measure dispersion.

Scales rank individuals, attitudes, or items along a continuum of the traits being scored.

The researcher knows the order of preference but not how much more one brand is preferred to another, i.e., the ordinal scale provides no information about the interval between any two brands. An ordinal scale offers all the information a nominal scale would have provided. Furthermore, positional statistics such as the median, quartile, and percentile can be calculated. With ranked data, order correlation can be tested. The two primary approaches, which will be explored later in the unit, are Spearman’s Ranked Correlation Coefficient and Kendall’s Coefficient of Concordance.

Because it solely serves the purpose of counting, numbers in the nominal scale can be interchanged. Ordinal scale numbers have meaning and cannot be interchanged.

• Students can be classified based on their A, B, C, D, E, F, and so on, with A being better than B. The grades are assigned in ascending order, from highest to lowest.
• At the university, teachers are classified as professors, associate professors, assistant professors, lecturers, etc.
• In good organisations, professionals are identified as GM, DGM, AGM, SR.MGR, MGR, Dy. MGR, Asst. MGR, and so on.
• A comparison of two or more households based on their annual income or expenditure, e.g.

Households	A	B	C	D	E
Annual Income (Rs)	5,000	9,000	7,000	13,000	21,000

If the highest income is given #1, then we write it as

Households	Order of Households based on Annual Income
A	E (1)
B	D(2)
C	B(3)
D	C(4)
E	A(5)

One can ask respondents questions based on one or more features, such as flower, colour, and so on, and inquire about likes or dislikes, such as whether the respondent likes soft beverages.

This method allows for obtaining a ranking by questioning the responder’s level of acceptance. The individual rankings can then be combined to get a group ranking.

The interval scale applies the concept of “equality of interval,” meaning it uses the intervals as the foundation for equating the units, assuming that the intervals are equal.

Only with interval-scaled data can researchers justify using the arithmetic mean as a measure of average. Because the interval or cardinal scale contains identical measurement units, it is possible to discern the order of scale scores and the distance between them. It should be noted, however, that the zero point on an interval scale is arbitrary and not a true zero. Of course, this has ramifications for the data manipulation and analysis that can be performed on data acquired in this format. It is possible to add or subtract a constant from all scale values without changing its shape, but the values cannot be multiplied or divided. Two respondents with scale positions 1 and 2 are as far apart as two respondents with scale positions 4 and 5, but a person with a score of 10 feels twice as strongly as one with a score of 5. Temperature is measured in Centigrade or Fahrenheit on an interval scale. We cannot say that 50°F is twice as hot as 25°F since the comparable centigrade temperatures, 100°C and -3.9°C, are not in the 2:1 ratio.

Interval scales can be numerical or conceptual.

Characteristics

• There is no absolute zero on interval scales. It is chosen at random.
• The mean is used to calculate central tendency.
• Standard deviation is used to calculate dispersion.
• The t-test and f-test are employed to determine significance.
• The scale is based on interval equality.

Use: The majority of basic statistical methods of analysis require interval scales. These are not recounted here because they are so ubiquitous and can be found in almost all introductory statistics textbooks.

5.1.3 Interval Scale

The interval scale outperforms the nominal and ordinal scales. The scale distance represents an equal distance on the property being measured. The interval scale can inform us “how far apart the items are about an attribute.” This means that the two can be compared. The difference between “1” and “2” is the same as the difference between “3” and “2.”

The interval scale applies the concept of “equality of interval,” meaning it uses the intervals as the foundation for equating the units, assuming that the intervals are equal.

Only with interval scale data can researchers justify using the arithmetic mean as a measure of average. Because the interval or cardinal scale contains identical measurement units, it is possible to discern the order of scale scores and the distance between them. It should be noted, however, that the zero point on an interval scale is arbitrary and not a true zero. Of course, this has ramifications for the data manipulation and analysis that can be performed on data acquired in this format. It is possible to add or subtract a constant from all scale values without changing its shape, but the values cannot be multiplied or divided. Two respondents with scale positions 1 and 2 are as far apart as two respondents with scale positions 4 and 5, but a person with a score of 10 feels twice as strongly as one with a score of 5. Temperature is measured in Centigrade or Fahrenheit on an interval scale. We cannot say that 50°F is twice as hot as 25°F since the comparable centigrade temperatures, 100°C and -3.9°C, are not in the 2:1 ratio.

Interval scales can be numerical or conceptual.

Characteristics

• There is no absolute zero on interval scales. It is chosen at random.
• The mean is used to calculate central tendency.
• Standard deviation is used to calculate dispersion.
• The t-test and f-test are employed to determine significance.
• The scale is based on interval equality.

Use: Most typical statistical methods of analysis require interval scales in order to be used. These are not recounted here because they are so ubiquitous and can be found in almost all introductory statistics textbooks.

Example:

1. Assume we want to rate a refrigerator using an interval scale. It will look like this:

(a) Brand Name                    Poor………..Good

(b) Price                                  High………..Low

(c) After-sales service          Poor……….Good

(d) Utility                               Poor……….Good

The researcher cannot deduce that a respondent with a rating of 6 is three times more positive about a product under consideration than another respondent with a rating of 2.

2. How many hours do you spend each day on class assignments?

(a) 30 minutes;

(b) 30 minutes to 1 hour

(c) 1 hour to 1-1/2 hours

(d) > 1-1/2 hours.

Implications for statistics: We can calculate the range, mean, median, etc.

5.1.4 Ratio Scale

A ratio scale is an internal scale with a meaningful zero point. It can be used to measure length, weight, or distance. Using this scale, it is possible to describe how many times greater or smaller one object is compared to the other.

These scales are used to quantify real-world characteristics. A ratio scale is the highest degree of measurement. This has the characteristics of an interval scale and a fixed origin or zero point. Weights, lengths, and times are examples of ratio-scaled variables. The researcher can use ratio scales to compare discrepancies in scores and the relative magnitude of scores. For example, the difference between 5 and 10 minutes is the same as the difference between 10 and 15 minutes, and 10 minutes is twice as long as 5 minutes.

Given that sociological and managerial research rarely strives beyond the interval level of measurement, it is not suggested that this level of analysis be given special attention. To summarise, almost all statistical processes can be done on ratio scales.

Characteristics

• This scale has a measurement of absolute zero.
• Geometric and harmonic means are utilised to calculate central tendency.

Use: The ratio scale can be applied to any statistical technique.

For example, sales of product A this year are twice as high as those of the identical product last year.

Statistical implications: This scale allows for the execution of all statistical operations.

5.2 Techniques of Developing Measurement Tools

Scale creation procedures assess a group’s or an individual’s attitude. In other words, the scale creation technique aids in estimating an individual’s or a group’s interest or behaviour toward another’s environment rather than oneself. When using a scale creation technique, you must consider several decisions relating to the attitude of the individual or group. Among these choices are determining the level of the included data and whether it is nominal, ordinal, interval, or ratio.

• It is identifying the most appropriate statistical analysis for scale creation.
• I am choosing the scale construction style to use.
• Choosing the physical architecture of the scales. Identifying the scale categories that must be employed.

There are two types of scale construction: comparative and non-comparative. The comparative technique is utilised by comparing the items to determine the scale values of several things. The non-comparative technique determines an object’s scale value without comparing it to another item. Furthermore, these two approaches come in a variety of varieties. The following are the numerous sorts of comparison techniques:

• Pairwise comparison scale: This is an ordinal level scale design technique in which a respondent is given two options and then asked to choose one.
• Rasch model scale: In this technique, numerous respondents are involved with several items simultaneously, and comparisons are derived from their responses to determine the scale values. The rate-order scale is another ordinal-level scale construction technique in which a respondent is given several things to rank.
• Constant sum scale: In this scale creation technique, a respondent is usually given a fixed amount of money, credits, or points to allocate to various objects to determine their scale values.

Non-comparative approaches are classified as follows:

• Continuous rating scale: In this technique, respondents generally rate an item using a sequence of scale points. This method is often referred to as visual rating scaling.
• Likert scale: This technique allows respondents to rate items on a five-to-seven-point scale based on how much they agree or disagree with the item.
• Semantic differential scale: In this technique, respondents are asked to score an item’s many qualities on a seven-point scale.

5.3 Scaling – Importance

Scaling is a technique or series of procedures used to assess an individual’s attitude. Scaling is described as the rule-based assignment of things to numbers. The definition’s objects are text declarations, which can be statements of attitude or principle. Scaling does not directly measure an individual’s attitude. It is initially migrated to statements, after which the numbers are assigned. The figure below depicts how to scale people’s attitudes.

In the graphic above, we will measure an individual’s attitude toward drinkers by analysing his thoughts about them. You can observe that as you go down, people’s attitudes or behaviours toward drinkers grow more ad hoc. If a person agrees with one of the statements in the list, he is more likely to agree with all of the assertions above that statement. As a result, the rule in this case is a growing one. This is referred to as scaling. Scaling is used in the research process to put the hypothesis to the test. Scaling can be used as part of a probing investigation at times.

5.4 Scaling Techniques (Comparative and Non-Comparative)

• Comparative Scales: A direct comparison of two or more objects.
• Non-comparative Scales: Objects or stimuli are scaled separately from one another.

5.4.1 Comparative Scaling Techniques

 Paired Comparison

Example

In this example, a responder is asked to indicate his flavour preferences among five coffee brands – A, B, C, D, and E. He must declare his preference in pairs. The number of pairs is computed as follows. The rated brands are displayed two at a time, so each brand in the category is compared to every other brand once. In each pair, respondents were asked to divide 100 points based on how much they preferred one. The score is unique to each brand.

N(N – 1)/2 = number of pairings

5(5 – 1)/2 in this example.

A&B	B&D
A&C	B&E
A&D	C&D
A&E	C&E
B&C	D&E

If we have 15 brands to analyse, we have 105 paired comparison(s), the method’s restriction.

For each pair of teachers, please indicate with a 1 which professor you prefer to take classes from.

	Cunningham	Day	Parker	Thomas
Cunningham		0	0	0
Day	1		1	0
Parker	1	0		0
Thomas	1	1	1	0
# of times Preferred	3	1	2	0

Rank Order Scaling 

• Respondents are shown multiple things at the same time.
• They are then asked to order or rate them based on some criterion.
• The collected data are ordinal—arranged or ranked in order of magnitude.
• It is commonly used to assess brand preferences and brand features.

For example, please rate the following instructors in order of preference. Assign a “1” to the instructor you prefer the most, a “2” to the instructor you prefer the second most, a “3” to the instructor you prefer the third most, and a “4” to the instructor you prefer the least.

Instructor	Ranking
Cunningham	1
Day	3
Parker	2
Thomas	4

Constant Sum Scaling

• Respondents are asked to assign a constant total of units to a group of stimulus objects based on some criterion.
• The units assigned represent the value assigned to the items.
• The data received are of the interval variety.
• Enables fine distinction between alternatives

For example, consider the four marketing professors listed below and the three factors that students often consider relevant. Please assign a number to each component that indicates how well you believe each instructor performs on that aspect. Greater numbers indicate higher scores. The sum of all the teachers’ scores on a certain aspect should equal 100.

Instructor	Availability	Fairness	Easy Tests
Cunningham	30	35	25
Day	30	25	25
Parker	25	25	25
Thomas	15	15	25
Sum Total	100	100	100

5.4.2  Non-comparative Scale

 Continuous Rating Scale

VERY POOR………………………………………………..VERY GOOD

0       10   20   30   40   50   60   70    80    90     100

The Likert Scale

It is referred to as a summated rating scale. This is made up of a succession of assertions about an attitude object. Each statement includes a ‘5-point scale with Agree and Disagree options. They are also known as summated scales since the results of individual items are added to generate a total score for the respondent. The Likert Scale is divided into two parts: the item part and the evaluation part. The item section comments about a particular product, event, or attitude. The evaluation section consists of responses ranging from “strongly agree” to “strongly disagree.” The five-point scale is employed in this case. Numbers such as +2, +1, 0, –1, –2 are utilised. Let us look at an example of how a customer’s opinion of a shopping mall is measured.

The overall attitude of the respondents is measured by adding (their) numerical rating on the statements that comprise the scale. Because some statements are favourable and others are unfavourable, completing them before summarising the ratings is critical. In other words, a “strongly agree” category is tied to a favourable statement, whereas a “strongly disagree” category is attached to an unfavourable one. The assertion must always be given the same numerical value, such as +2 or –2. The Likert Scale’s success is determined by “how successfully the assertions are generated?” The more favourable the attitude, the higher the respondent’s score. For example, if there are two shopping malls, ABC and XYZ, and the scores on the Likert Scale are 30 and 60, respectively, we can conclude that people prefer XYZ over ABC.

Differential Semantic Scale

This is comparable to the Likert Scale. It also includes several items for respondents to rate. The primary distinction between Likert and Semantic Differential Scale is as follows:

It employs “Bipolar” adverbs and phrases. The Semantic Differential Scale has no statements.

A seven-point scale separates each pair of adjectives.

Items on the Semantic Differential Scale

Please rate the five real estate developers listed below using the scales provided for each factor. Developers are involved.

S. No.	Scale items	–3	–2	–1	0	+1	+2	+3	–
1.	Not reliable	_	_	_	_	_	_	_	Reliable
2.	Expensive	_	_	_	_	_	_	_	Not expensive
3.	Trustworthy	_	_	_	_	_	_	_	Not trustworthy
4.	Untimely delivery	_	_	_	_	_	_	_	Timely delivery
5.	Strong Brand Image	_	_	_	_	_	_	_	Poor brand image

The respondents were asked to select one of seven categories that best described their attitudes. The computation is carried out in the same manner as in the Likert Scale. Assume we are attempting to assess the packaging of a specific product. The seven-point scale will look like this:

“I’m feeling…………..

• Delighted
• Pleased
• Mostly satisfied
• Equally satisfied and dissatisfied
• Mostly dissatisfied
• Unhappy

The Thurstone Scale

This is also known as a scale with equal-looking intervals. The steps for building a Thurstone Scale are as follows:

• Step 1: Make many remarks about the attitude to be measured.
• Step 2: These assertions (75 to 100) are handed to a panel of judges, say 20 to 30, who are asked to categorise them based on their favorability and unfavorability.
• Step 3: The judges must create 11 piles. The piles range from “most unfavourable” in pile 1 to “neutral” in pile 6 and “most favourable” in pile 11.
• Step 4: Examine the frequency distribution of ratings for each statement and delete statements with highly disparate ratings from various judges.
• Step 5: Choose one or two statements from each of the 11 piles for the final scale. To create the scale, list the selected statements in random order.
• Step 6: The respondents whose attitudes were to be scaled were given a set of statements and asked to indicate whether they agreed or disagreed with each one. Some may agree with only one statement, while others may agree with multiple statements.

Example:

1. Films on crime and violence:

• All films depicting crime and violence should be illegal.
• Watching movies about crime and violence is a waste of time.
• The majority of crime films are harmful and destructive.
• The majority of criminal films are monotonous in terms of direction and theme.
• Watching a crime and violent film does not disrupt my daily routine.
• I have no strong feelings about watching crime and violence-themed films.
• I enjoy watching crime and violent films.
• The majority of crime and violence films are exciting and engaging.
• Crime films serve as an information bank for the spectator.
• Viewing a crime film teaches people “how to be safe and protect oneself.”
• Seeing criminality in a movie does not impair our way of life.

Conclusion: Statements 8, 9, and 10 may be agreed upon by a response. This type of agreement demonstrates a favourable attitude toward crime and violence. On the other hand, if respondents choose items 1, 3, and 4, they are unfavourably disposed toward crime in movies. If the respondent selects 1, 5, and 11, this could indicate that he or she is inconsistent in his or her attitude about the subject.

2. Assume we’re interested in the attitudes of a specific socioeconomic group of respondents toward savings and investments. The following would be the final set of statements:

• One should live in the now rather than the future. Savings are, therefore,, not necessary.
• There are numerous attractions where the money saved can be spent.
• It is preferable to spend savings rather than risk them in investments.
• Investments are risky because the money is likewise frozen.
• You earn to spend rather than to invest.
• There is no way to save these days.
• A portion of one’s earnings should be saved and invested.
• The future is unclear, but investments will keep us safe.
• Every individual should have some savings and investments.
• One should endeavour to save more money so that most of it can be invested.
• All savings should be put towards investments for the future.

Conclusion: A respondent who agrees with points 8, 9, and 11 is thought to have a positive attitude regarding saving and investing. The individual who agrees with propositions 2, 3, and 4 has a negative attitude. Furthermore, a respondent’s attitude is inconsistent if he selects statements 1, 3, 7, or 9.

Multidimensional Scaling

This is used to research customer attitudes, specifically perceptions and preferences. These strategies aid in identifying and quantifying the product features that are essential to customers. Multi-Dimensional Scaling can be used to investigate the following topics:

• What are the most important factors to consider when selecting a product (soft drinks, modes of transportation)?
• When comparing different product brands, which characteristics do customers compare? Is it the price, the quality, the availability, and so on?
• What is the best mix of attributes, according to the consumer? (That is, which two or more attributes will the buyer consider before making a purchase decision?)
• Which advertising messages are compatible with the consumer’s impressions of the brand?

There are two methods for gathering input data for perceptual mapping:

• Non-attribute technique: In this method, the researcher asks the respondent to judge the objects. The respondent determines the criteria for comparing the objects in this method.
• Attribute method: Instead of respondents choosing the criteria, they were asked to compare the objects using the criteria supplied by the researcher.

To determine a consumer’s perception, for example, Assume there are five insurance companies to be evaluated based on two factors:

(1) convenient location and

(2) pleasant personal service.

The following are customers’ perceptions of the five insurance companies:

Five insurance firms are A, B, C, D, and E.

• According to the map, B and E are two different insurance companies.
• C is in a very convenient location.
• A is in a less convenient location than E.
• D is less convenient in terms of location than C.
• E is a less convenient location than D.

MDS uses software such as SPSS, SAS, and Excel. One of SPSS’s key characteristics is brand positioning research. SAS is a business intelligence application. Excel is also used to some extent.

Stapel Scales

• When it is challenging to generate pairs of bipolar adjectives, modern versions of the Stapel scale substitute a single adjective for the semantic differential.
• The benefits and drawbacks of a Stapel scale and the outcomes are quite comparable to those of a semantic differential.

The Stapel and other scales are easier to conduct and administer.

5.5 Acceptance Criteria for a Good Test

There are two factors for determining whether the scale used is good. They are as follows: 1. Reliability; and 2. Validity.

5.5.1 Reliability Analysis

The degree to which the measurement method is error-free is called reliability. Reliability is concerned with precision and consistency. The scale is trustworthy if repeated measurements under constant conditions provide the same findings.

For example, consider your attitude toward a product or brand preference.

The same scale can be used on the same group of respondents in the same manner to assure reliability. However, in practice, this isn’t easy because:

• The extent to which a scale yields consistent findings.
• Test-retest Reliability: Scales are administered to respondents twice under substantially identical conditions.
• Alternative-form Reliability: Two equivalent scale forms are built and evaluated with the same respondents twice.
• Internal Coherence (a) The consistency with which each item represents the construct of interest (b) Used to assess the dependability of a summated scale (c) Split-half Reliability
• Scale items are divided into two halves, and the resulting half scores are correlated: Alpha coefficient (most common test of reliability)
• Average of all possible split-half coefficients resulting from varied scale item splitting.

5.5.2 Validity Analysis

The validity paradigm was centred on the question, “Are we measuring what we think we’re measuring?” The scale’s success depends on measuring “What is supposed to be measured?” The most important of the two scaling attributes is validity.

There are numerous methods for determining the accuracy of the scale used for measurement:

1. Construct Validity: A sales manager feels that there is a direct relationship between job satisfaction and the degree to which a person is an extrovert, as well as the work performance of his sales team. As a result, those with high job satisfaction and outgoing personalities should perform well. If they don’t, we can question the measure’s construct validity.

2. Content Validity: A researcher should adequately describe the problem. Determine the item to be measured. Create a scale that is appropriate for this purpose. Despite these advantages, the scale may be criticised for lacking content validity. Face validity is another term for content validity. The advent of new packaged foods is one example. When a new packaged food product is introduced, the product represents a significant change in flavour. Thousands of people could be asked to sample the new packed meal. People may report that they liked the new flavour overwhelmingly. Even with such a positive response, the product may fail when it is presented on a commercial basis. So, what’s the problem? Perhaps a critical question was left out. People were asked if they liked the new packaged food, and the majority said “yes,” but they were not asked, “Are you willing to give up the product that you are now consuming?” The fault was not clearly stated in this example, and the item to be ‘measured’ was omitted.

3. Predictive Validity refers to “how well a researcher can forecast future performance based on knowledge of attitude scores.”

For example, an opinion questionnaire that forecasts product demand has predictive validity. The predictive validity approach measures attitude first and then forecasts future behaviour. Finally, future behaviour is calculated at the proper time. Compare the two outcomes (past and future). The scale has predictive validity if the two scores are closely related.

4. Criterion Validity:

(a) Determines whether the measuring scale performs as expected in connection to other variables chosen as meaningful criteria, i.e., predicted and actual behaviour should be comparable.

(b) Addresses the issue of what construct or attribute the scale is measuring.

5. Convergent Validity: The extent to which a scale corresponds with other measures of the same construct.

6. Discriminant Validity: The extent to which a measure does not connect with other constructs that are meant to differ from it.

7. Nomological Validity: The extent to which a scale correlates with assessments of various but related constructs in theoretically predictable ways.

REVIEW QUESTIONS:

1. What are the merits of the Thurstone Scale in your analysis?
2. What limitations might you encounter with the Thurstone Scale?
3. Which method do you consider more favourable between attribute and non-attribute perceptual mapping, and why?
4. From your perspective, what are the potential uses of multi-dimensional scaling?
5. What other major drawbacks do you see with MDS besides its potential variability over time?
6. What factors contribute to the difficulty in maintaining reliability in measurement?
7. Does a measurement scale consistently perform as expected concerning other selected variables as meaningful criteria? Why or why not?
8. On average, how many cups of tea do you consume in a day, and what factors influence your response? Provide a technical explanation.
9. Describe the construction of the (a) Likert scale, (b) Semantic differential scale, and (c) Thurstone scale.
10. Despite being reliable, a scale may lack content validity. Discuss.
11. Identify the type of scale you would use in each of the following scenarios (ordinal, nominal, interval, ratio), and justify your choice.